Последние изменения в многократной склерозовой нейропатологии.

INTRODUCTION

Multiple sclerosis (MS) is a chronic inflammatory central nervous system (CNS) disease with characteristic tissue features. In the last decades, clinical, pathological, and experimental MS research has developed important concepts that led to substantial therapeutic advances. Nowadays, people with MS (pwMS) face substantially fewer relapses and reach disability milestones considerably later. A plethora of anti-inflammatory treatments now make it possible to tailor the therapeutic approach specifically to the needs of the individual patient, making the therapeutic advances in MS an outstanding success story in clinical neurology [ 1 ].

no caption available

Yet even the most effective anti-inflammatory treatments do not lead to complete disease resolution, and the concept of disability accumulation without apparent inflammatory activity (progression independent of relapse activity – PIRA) and its underlying pathogenetic mechanisms are intensely discussed [ 2 , 3 ]. Thus, MS researchers rely heavily on neuropathological and neuroimaging findings. To better understand and treat disability progression in MS, recent neuropathological research has focused on two aspects: inflammatory processes that might escape visualization by ‘conventional’ MR imaging, and identification of biomarkers and mechanisms of neurodegeneration [ 4 , 5 ]. Many recent studies have, therefore, focused on myeloid cell activation in individuals with chronic progressive disease and its relation to lesion evolution and neurodegeneration. As the precise etiopathogenesis of MS is unknown, recent neuropathological studies have also addressed mechanisms of initial lesion formation and lesion evolution. Progress has also been made in our understanding of related idiopathic inflammatory demyelinating diseases (IIDD), in particular, MOG antibody-associated disease (MOGAD) and neuromyelitis optica spectrum disorder (NMOSD). Biomarkers that capture the underlying pathogenetic principles of individual IIDDs are currently being validated and will help to distinguish MS from other IIDDs [ 6 ].

Recent technical advancements in human tissue studies have greatly enhanced our ability to obtain molecular data from archival human samples. Driven by significant advances in sequencing and bioinformatics technologies, a much deeper and unbiased characterization of cellular transcriptomes, both at the single-cell level following tissue dissociation and within the intact tissue, is now possible [ 7 ]. This marks a turning point in tissue-based research and a crucial first step towards a deeper understanding of cellular function in both tissue and disease contexts. In MS, these ‘spatial transcriptome’ technologies allow the identification of cellular identities and transcriptome-inferred function in different types of lesions in relation to the surrounding nonlesioned tissue [ 8 ].

In the following, we will present recent advances in MS neuropathology, highlight novel concepts from these studies, and discuss their potential to improve care for pwMS.

CLINICOPATHOLOGICAL CORRELATION OF MULTIPLE SCLEROSIS: NEW CONCEPTS AND BASIC PRINCIPLES REVISITED

During the last decades, large MS autoptic cohorts have been collected that allow now to study the association between specific pathological features and clinical parameters. Two recent studies building on tissue from the UK MS Brain Bank focused on the interplay of inflammation, demyelination, and neurodegeneration. In a large retrospective postmortem cohort study of 269 patients, Nicholas et al. [ 9 ▪ ] could demonstrate that active demyelinating lesions as well as substantial perivenous lymphocyte infiltration can still be found at the time of death in a subgroup of patients. These features correlated with high levels of meningeal inflammation and were linked to rapid disease progression, indicating inflammation as a driver of progression. Interestingly, in a study comparing neuropathological features and clinical characteristics of people with early disease onset to people with late disease onset, neuron density in the thalamus and pons, but not active demyelination or compartmentalized inflammation, was most associated with disability outcomes in the late onset cohort [ 10 ]. These studies thus suggest a link between inflammation, demyelination, and neurodegeneration that may depend on further variables, such as genetics, environment, age, and others. They also call for biomarkers to stratify patients according to their predominant tissue phenotype.

In further studies investigating the link between demyelination and neurodegeneration, the term ‘myelocortical MS’ was coined to describe a disease pattern in which cortical demyelination and neuron loss as well as spinal cord pathology predominate, while cerebral white matter demyelination is nearly absent. This pattern was clinically associated with a primary progressive disease phenotype [ 11 ]. In a recent correlative postmortem MRI and pathology study, axonal damage and density reduction around the cerebral ventricles in the absence of overt demyelination were identified, further strengthening the above concept and supporting surface-in gradients of neurodegeneration in MS [ 12 ].

In a study of 119 patients, Waldman et al. [ 13 ] documented a high proportion of inflammatory demyelinated lesions in the cord, when assessing myeloid cell density and activation. They observed a propensity of the cervical cord for lesion formation, and an association of lesion load and activity with clinical disease milestones. The authors observe a relative sparing of subpial areas and find that intraspinal lesions relate to vascular topography.

In another approach to establish a meaningful relationship between neuropathological features obtained at postmortem with clinical and genetic data, a large study from the Amsterdam MS Brain Bank with data from 226 MS donors identified three dimensions for grouping patients according to, for example, lesion phenotypes, cortical pathology, and myeloid cell reactivity. The dimensions were interpreted to represent mechanisms of disease progression, lesion initiation, and loss of lesion activity and scar formation. Their work may serve to disentangle MS clinical and lesion heterogeneity and translate into living cohorts with MS [ 14 ].

Taken together, large neuropathological cohort studies on postmortem MS tissue support the notion that adaptive inflammation continues to associate with disease progression throughout the disease course, and that spinal lesion load and neuroaxonal pathology contribute substantially to disability outcomes.

TOWARDS A MOLECULAR UNDERSTANDING OF LESION FORMATION AND EVOLUTION

Several ‘spatial transcriptome’ studies on MS postmortem brains were recently published that used different technologies and in part reached single cell resolution. They all could identify classic neuropathological tissue landmarks, such as cortical layers, vessels, and MS lesion borders. However, with bioinformatic clustering methods to group cells with similar expression profiles, these studies could also reveal lesion areas and changes in the so-called normal-appearing white matter that are not visualized by conventional tissue staining methods. Lerma-Martin et al. [ 15 ▪▪ ] in their study highlight the concept of so-called ‘tissue niches’ in MS brains, for example, the perivascular space, the inflamed lesion rim, or the lesion core, each harboring cells with specific transcriptional profiles and defined cellular interactions. In a technically similar approach, Alsema et al. [ 16 ▪▪ ] generated spatial expression maps of white and grey matter MS lesions and identified white matter lesion progression trajectories, thereby defining molecular correlates of how normal-appearing white matter may develop into active or mixed active/inactive lesions. Kukanja et al. [ 17 ▪▪ ] examined MS spinal cord lesions alongside tissue of mice with experimental autoimmune encephalomyelitis (EAE) and reached the single cell level. Their study identifies the molecular determinants of lesion progression and resolution (in mice) and is the first to compare oligodendroglial and astroglial phenotypes between mice and humans in a spatial setting. These studies highlight the complex cellular interactions in MS lesions and provide important resources, for example, for plausibility testing of novel theories and therapeutic targets related to MS lesion evolution.

REVEALING PATHOLOGICAL CORRELATES OF DISEASE PROGRESSION

Measuring cortical disease activity

A major focus of current clinical and pathological MS research is the identification of cellular and molecular correlates of disease progression, with the goal to identify meaningful targets for therapeutic intervention. Previous studies highlighted cortical demyelinated lesions as a disease manifestation strongly associated with the chronic progressive disease stage [ 18 , 19 ]. However, cortical demyelinated lesions are notoriously difficult to visualize by MRI in pwMS [ 20 ]. In a recent study, Okutan et al. [ 21 ▪ ] identified a potential surrogate of intracortical lesions. They observed that subcortical/juxtacortical lesions that are readily visible on standard MRI correlate with cortical lesions, cortical lesion activity, and meningeal perivascular inflammation, thus potentially representing a useful surrogate of corticomeningeal inflammatory pathology. In a combined postmortem MRI/pathology study, a specific variant of these lesions, juxtacortical paramagnetic rim lesions, were histologically characterized by iron-containing foam cells. In in-vivo cohorts, these lesions identified pwMS with higher cortical lesion load, supporting their role as a potential biomarker of cortical pathology [ 22 ] (Fig. 1 ).

Cortical disease activity in progressive multiple sclerosis. Iba1+ microglia cells in a subpial cortical demyelinated area (a and c). MHCII immunohistochemistry (IHC) indicates pronounced microglia activation at the border of the demyelinated upper cortical layers to the adjacent regularly myelinated tissue (b and d). Frame (a) is magnified in (c). Double immunofluorescent IHC (a and c) for myelin basic protein (MBP) and the myeloid cell marker Iba1 (Iba1: yellow; MBP: cyclamen); light microscopic double-labellings (b and d) for MBP (blue) and MHCII (brown).

The high interest in imaging cortical and meningeal inflammation as a surrogate of disease progression is also reflected by the study of Herranz et al. [ 23 ▪ ] demonstrating that myeloid cell activation in the meninges and upper cortical layers can be measured by PET imaging for 18 kDa mitochondrial translocator protein (TSPO). PET TSPO levels correlated with the EDSS, and TSPO was detected by immunohistochemistry on meningeal and cortical activated MHC II-positive myeloid cells, including Tmem119-positive microglia. Another recent study on cortical microglia activation examined the microglia rim of subpial cortical lesions in comparison to the respective lesion rim of white matter lesions. The authors found distinct myeloid cell morphologies and identified distinct transcription variations between microglia lining grey and white matter lesions, with white matter microglia demonstrating higher expression levels of immune-related transcripts. Also, transcripts involved in iron homeostasis and phagocytosis were much more prevalent in white matter than in cortical lesion rim myeloid cells [ 24 ].

Myeloid cell activation, polarization, and therapeutic targetability

Several neuropathological studies focused on microglia activation as a correlate of insidious, ‘non inflammatory’ disease progression. Microglia are yolk sac-derived myeloid cells that invade the CNS early in ontogenesis and co-evolve as the intraparenchymal innate immune cells of the CNS. Phenotypic and molecular signs of microglia activation are observed in all CNS diseases, and their beneficial or detrimental role is subject to controversy. The contribution of microglia to myelin phagocytosis in MS is also not entirely resolved, and the walling off of evolving demyelinating lesions may suggest a protective rather than harmful phenotype [ 25 ]. However, microglia activation associates with early (re)active lesion phenotypes in MS, and an in-depth study of microglial nodules revealed inflammatory polarization in MS compared to stroke [ 26 ▪ ]. In a recent study, no correlation of microglia density with the density of mature oligodendrocytes was observed in the periplaque white matter [ 27 ]. However, microglia activation was more pronounced in younger pwMS with more mixed active lesions. Microglia are important contributors to the myeloid cell rim of chronic active MS lesions [ 28 ]. Of interest, a subtype of these lesions demonstrated specific imaging properties in MRI, the so-called ‘paramagnetic rim lesions’ (see also above). Histologically, they are characterized by an accumulation of iron and iron-metabolizing enzymes in myeloid cells [ 29 ▪▪ ]. Their presence correlates with faster disability accrual and higher serum neurofilament levels [ 30 ]. In line with current therapeutic concepts, increased BTK expression was found in iron-bearing myeloid cells, highlighting their inflammatory activation and potential targetability [ 31 ].

MYELIN AND OLIGODENDROCYTE DAMAGE AND REGENERATION

In an attempt to identify pathological correlates of very early MS lesion formation, Luchicchi et al. [ 32 , 33 ] focused on micro-diffusely abnormal white matter characterized by myelin blisters, degradation of myelin basic protein, increase in polar myelin phospholipids and citrullination but no evidence of Wallerian degeneration. Whereas oligodendrocytes are abundant in early resorptive MS lesions, later lesion stages suffer from a decrease in mature oligodendrocytes and oligodendrocyte precursor cells [ 34 , 35 ]. To better understand the mechanisms of oligodendrocyte demise and remyelination failure, two recent studies determined markers of cellular senescence in lesions and observed an upregulation in oligodendrocytes as well as astrocytes, neurons, endothelia, and myeloid cells [ 36 , 37 ]. Cellular senescence correlated with lesion activity and associated with faster disability progression.

Remyelination is a well orchestrated regenerative process by which denuded axons are re-installed with functional myelin sheaths [ 38 ]. Given the insulating and nutritive properties of the myelin sheath, remyelination is considered axonoprotective, and pro-remyelinating therapies are tested for progressive disease [ 39 ]. Remyelination efficacy is highly variable among pwMS [ 40 ]. Comparing different lesion types of efficient and poor remyelinators by bulk sequencing, both known and novel transcripts related to oligdendrocyte differentiation and remyelination were identified [ 41 ▪▪ ]. Also, iNOS + pro-inflammatory myeloid cells positively correlated with nonprocess-bearing BCAS1 + oligodendrocytes, potentially reflecting impaired differentiation [ 42 ].

Recently developed imaging modalities, including PET and advanced magnetic resonance (MR) imaging, allow for the detection of remyelination in vivo [ 43 ]. However, the specificity to myelin of certain MR sequences is still discussed [ 44 ]. In a whole brain pathology/MRI study, both myelin water imaging modalities applied revealed high specificity and sensitivity to detect even a few remyelinated fibers, as demonstrated by corresponding immunohistochemistry [ 45 ]. Also adding to the complexity of MRI-based tissue analysis, a recent study revealed a striking presence of astrocytes in remyelinated lesions, which may aid in determining the temporal evolution of remyelinated shadow plaques and in defining molecular conditions that support extensive remyelination [ 46 ].

Interestingly, the cortical grey matter is not only a hot spot of demyelination and microglia activation but also represents an environment particularly conducive to remyelination [ 47 , 48 ] (Fig. 2 ). Recent findings demonstrated that perineuronal satellite oligodendrocytes contribute to remyelination in MS and experimental models of demyelination [ 49 ▪ ]. In pwMS, 3T MRI imaging revealed that cortical demyelination relates to disability accrual whereas early cortical remyelination reduced the risk of disease progression [ 50 ]. These studies underline that the cortical grey matter is a disease-determining site in MS.

Ongoing and established cortical remyelination. BCAS1 IHC (a–d) identifies myelin sheaths and process-bearing oligodendrocytes indicative of ongoing remyelination (b). Established cortical remyelination is characterized by a reduced density of thin myelin sheaths (d). IHC, immunohistochemistry; NAC, normal-appearing cortex (a and c); RM, remyelination.

Whereas most studies conceptualize myelin as axonoprotective [ 39 ], a recent study combining electron microscopy in human biopsy samples and data from experimental models reports that axonal damage was mainly observed in still myelinated axons, raising the question whether myelin itself may in fact predispose axons for damage under autoimmune conditions. The presence of relevant antigens in the myelin sheath as well as a dependence of myelinated axons on nutritional support by oligodendrocytes might contribute to this finding [ 51 ].

CONCLUSION

In recent years, MS neuropathology has strongly focused on tissue correlates of disease progression, in particular microglia activation, neuroaxonal pathology, and cortical demyelinated lesions. Importantly, many studies support a role of adaptive inflammation, in particular in a perivenous and meningeal localization, as a correlate of severe and rapidly progressive disease also in patients with long disease duration. One crucial goal will be to identify these highly inflammatory patients and stratify them for effective anti-inflammatory therapies. The development of respective imaging and fluid biomarkers is highly relevant and supported by neuropathological studies. Also, the advent of novel molecular technologies applicable to human CNS tissue provides valuable resources for investigating lesion formation and evolution. Upcoming spatially resolved proteome technologies will further complement the molecular analysis of MS patient tissues. These technologies with their comprehensive and unbiased approach will aid in defining novel hubs for inflammation and tissue destruction, and thus support the development of therapies specifically targeting the intersection of inflammation and neuroaxonal damage.

Acknowledgements

We thank Cynthia Bunker for language editing.

Financial support and sponsorship

This work was supported by the Deutsche Forschungsgemeinschaft (DFG) transregional collaborative research center TRR274/2, Project ID 408885537 ‘Checkpoints of CNS recovery’, STA 1389/5-1, NE 2447/1-1, the DFG Priority Programme ‘Local and Peripheral Drivers of Microglial Diversity and Function’ (SPP 2395; NE 2447/2–1; STA 1389/7-1), the DFG under Germany's Excellence Strategy (EXC 2067/1-390729940), the Gemeinnützige Hertie Foundation, the Deutsche Multiple Sklerose Gesellschaft (DMSG), and the National MS Society (USA). J.F. was supported by the UMG clinician scientist program ‘Cellular dynamics in pathogenesis and therapy’.

Conflicts of interest

There are no conflicts of interest.